PTAB

IPR2025-01248

Meta Platform Inc v. Sterling Computers Corp

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: System for determining a relevance score of electronic content
  • Brief Description: The ’217 patent describes a computer-implemented system for assigning relevance scores to electronic content, such as emails. The system uses several "computer program modules" to monitor a user’s actions, determine a first measure of relevance based on those actions, identify a "cognate" group of other users, determine a second measure of relevance based on factors from that group, and combine the measures to calculate a final relevance score.

3. Grounds for Unpatentability

Ground 1: Claims 1-3, 5-9, 13-15, 18-19, and 22 are obvious over Dumais

  • Prior Art Relied Upon: Dumais (Application # 2007/0016553A1).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Dumais, which discloses a data processing architecture using user interaction data, taught every limitation of the independent claims. Dumais’s system monitored user interactions (e.g., opening, editing files) and used a "tag component" to assign weights or scores, meeting the "first measure of relevance" based on user actions. Dumais also employed "collaborative filtering techniques" to analyze similarities among groups of users and develop user profiles, which Petitioner mapped to identifying "cognate" users and determining a "second measure of relevance." Finally, Dumais combined these factors to influence the rank of search results, which Petitioner contended was equivalent to determining the claimed "relevance score."
    • Motivation to Combine (for §103 grounds): Not applicable (single reference).
    • Expectation of Success (for §103 grounds): Not applicable (single reference).

Ground 2: Claims 1-22 are obvious over Dumais in view of Marston

  • Prior Art Relied Upon: Dumais (Application # 2007/0016553A1) and Marston (Application # 2004/0260710).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground combined the relevance scoring framework of Dumais with the specific "relational mail architecture" taught by Marston. Petitioner argued Marston’s architecture, where each message and sub-message is stored as a discrete, individually referenced object, disclosed a structure for the claimed "monitoring module." This granular structure inherently allowed for the detailed tracking of user actions required by the claims. The remaining relevance-scoring and user-grouping limitations were argued to be taught by Dumais, as in Ground 1.
    • Motivation to Combine (for §103 grounds): A POSITA would combine Marston’s efficient architecture with Dumais’s advanced filtering system to gain known benefits. Marston’s system reduced infrastructure costs by eliminating duplicate emails, while Dumais’s system addressed the proliferation of spam and irrelevant content. The combination was presented as a simple and predictable substitution of known, analogous art.
    • Expectation of Success (for §103 grounds): Petitioner asserted a high expectation of success, as the combination involved integrating two known types of email management systems to achieve a predictable result.

Ground 3: Claims 1-3, 5-9, 13-15, and 18-22 are obvious over Kircher in view of Krug

  • Prior Art Relied Upon: Kircher (Application # 2003/0195937A1) and Krug (WIPO Patent 2003/023650A2).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground combined Kircher's email prioritization system with Krug's user and object profiling techniques. Kircher taught monitoring a user's response to messages and ranking them based on criteria like sender importance, satisfying the "first measure of relevance." Krug taught building user profiles based on the "usage behavior" of individuals and groups, calculating "action equivalence" for user groups, and determining a final "comparison value" (relevance score). Petitioner argued Krug’s group-based analysis, which included user groups accessing a specific object, taught identifying cognate users and determining the "second measure of relevance."
    • Motivation to Combine (for §103 grounds): A POSITA would combine the references to improve the accuracy of relevance calculations in Kircher’s system by incorporating the more sophisticated collaborative filtering methods from Krug. Both references aimed to solve the same problems of spam and information overload. Combining multiple well-known filtering techniques was a common and predictable approach for improving system performance.
    • Expectation of Success (for §103 grounds): Petitioner argued a POSITA would have had a high likelihood of success because Kircher and Krug disclose analogous functionality, and combining their respective non-redundant features to obtain the benefits of both was a straightforward design choice.

  • Additional Grounds: Petitioner asserted an additional challenge based on Dumais in view of Applicant Admitted Prior Art (AAPA). A further ground asserted that claims 1, 4, 10-12, 16-17, and 20-21 are obvious over the combination of Kircher and Krug in view of Marston.

4. Key Claim Construction Positions

  • Petitioner argued that the four "computer program module" terms recited in independent claims 1 and 6 (e.g., "monitoring module," "first relevance measurement module") were subject to 35 U.S.C. §112, ¶ 6 as means-plus-function limitations.
  • Petitioner contended that "module" is a well-known nonce word that fails to recite a sufficiently definite structure for performing the claimed functions. As such, the claims should be limited to the corresponding algorithms disclosed in the ’217 patent’s specification. This construction was central to Petitioner’s argument, as it defined the specific structures that the prior art must be shown to disclose.

5. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1-22 of the ’217 patent as unpatentable.